In The prosody of formulaic sequences: A corpus and discourse approach, Phoebe M. S. Lin presents a remarkably comprehensive review of the state of knowledge about formulaic language and its prosodic features, and also presents three innovative studies of various aspects thereof. This book is a must-read for anyone working with formulaic language, especially those grappling with the challenging task of trying to determine how to identify formulaic language in smaller corpora and in spoken corpora. It also provides plenty of useful knowledge of features of formulaic language for those involved in any mode of research on formulaic language, and those looking to integrate current knowledge of formulaic language in language teaching.
Formulaic sequences are generally defined as multiword units of language which co-occur regularly and which tend to have a unitary meaning or function (Wood Reference Wood2015). They may also be prefabricated or stored and retrieved mentally as if they were a single word, although this aspect is still under scrutiny, particularly by psycholinguists. Formulaic sequences appear to display particular phonological characteristics in speech, and Lin explores these characteristics.
Not until this book has the work on the prosody of formulaic language been pulled together, although it has been a focus of research for some time. Phonological coherence has been a term frequently used for a key aspect of the nature of the prosody of formulaic sequences in spoken language. Peters (Reference Peters1977; Reference Peters1983) was among the first to observe that children often utter sequences which outstrip their grammatical competence, and which tend to contain no internal hesitations and be characterized by a smooth intonation contour, standing out from the rest of the flow of speech. Child language researchers such as Hickey (Reference Hickey1993) have noted that phonological coherence is basic to formulaic sequences in children's speech. As for adult speech, numerous researchers have assumed that a similar dynamic applies as to child language (e.g. Wray Reference Wray2002).
Some evidence points to phonological coherence being a function of holistic mental storage and retrieval of formulaic sequences, possibly in addition to frequency effects. It has been shown that high-frequency phrases undergo phonological reduction more quickly than other phrases (e.g. Bybee Reference Bybee2002). This is apparent in, for example, reduced schwa and sometimes t/d deletion (Bybee Reference Bybee, Barlow and Kemmer2000). Other research investigates pre-utterance pausing and intonation. Some work shows that pausing before formulaic sequences in spontaneous speech is not the same as before non-formulaic language. For example, Erman (Reference Erman2007) noted that retrieval of formulaic language appears to occur with shorter pauses before utterance. Some research has also uncovered an alignment between formulaic sequences with intonation units. For example, Lin & Adolphs (Reference Lin, Adolphs, Barfield and Gyllstad2009) revealed that the sequence ‘I don't know why’, taken from the Nottingham Corpus of Learner English, was uttered as a single intonation unit over half of the time. As well, Lin (Reference Lin and Wood2010) found that 82 per cent of sequences in a university lecture aligned with intonation boundaries on one side of the sequence, and 40 per cent aligned on both sides.
Lin's book contains seven densely packed chapters starting with an introduction, then an overview of formulaic language research, followed by a consideration of the idea that identification of formulaic language can be guided by prosodic cues. Lin then presents three studies, each in their own chapter: alignment of formulaic sequences with intonation units; a comprehensive profile of the intonation, stress and rhythm of formulaic language; and a multimodal approach to the identification of formulaic language by native-speaker judgement. The volume concludes with a summary of the knowledge presented and a look at the implications and the future of such research.
The first two chapters of the book provide a wealth of background on formulaic language and its identification in discourse. The introductory chapter lays out a definition of formulaic language and the structure of the following sections. The next chapter is an overview of issues concerning the definition and identification of formulaic language, laying a foundation for the three subsequent chapters which present research studies. Using Hudson & Wiktorsson's (Reference Hudson, Wiktorsson, Corrigan, Moravcsik, Ouali and Wheatley2009: 81) definition of formulaic language as ‘any sequence of two or more words that are perceived to be more constrained than usual in their co-occurrence’, Lin digs deeply into the historical records to show that formulaic language is, in fact, an umbrella term used to cover a large number of subtypes which are described and discussed using vague and overlapping terms. She goes on to note that child language researchers were among the first to elaborate criteria for formulaic language, covering a range of linguistic aspects. The discussion boils down to two main methods of identification, corpus-based extraction and native-speaker judgement. Lin goes on to elaborate on the nature of native-speaker judgement, drawing on previous work by Wulff (Reference Wulff2008) which shows that even lay judges draw subconsciously on knowledge of factors such as frequency and noncompositionality to determine what is formulaic. She considers measures needed to counteract judgement fatigue and confusion, including clear instructions and careful pilot testing of procedures, and also tackles how to define formulaic language for native speaker judges. Lin presents the criteria used by researchers in the past, and takes care to consider the fact that some of them used expert judges rather than lay judges. In the end, it appears that using simple, everyday language and appropriate examples with lay judges is a potentially fruitful means of identifying formulaic language. The entire discussion in chapter 2 is essential reading for anyone attempting to embark on a research project on formulaic language using judgement procedures, and indeed, for anyone planning to read the literature with a critical, informed eye.
The book's third chapter delves more deeply into whether prosodic cues are an aid to the identification of formulaic language, examining the literature on the speech of children, language learners and native speakers. Lin notes that issues with the results of corpus-based extraction research and native-speaker judgement research may be alleviated by using prosodic cues, which she calls the ‘phonological method’. She tracks the evolution of this method in research, noting that there is little clear evidence that formulaic language in adult speech can be identified using only prosodic cues. She also explores the idea that prosodic cues are an indicator of holistic processing of formulaic language, and lists the cues which are generally associated with formulaic language: intonation, pauses and pause phenomena, and stress placement. All of these inform the methods used in the three studies which are reported in subsequent chapters. The volume of literature covered in this chapter and the fresh conclusions drawn from the review are not to be found anywhere else. Again, this is essential reading for those working in identifying formulaic language in spoken language.
Chapter 4 presents the first of three studies examining whether formulaic sequences align with intonation units. Reviewing literature on intonation unit boundaries and spontaneous speech, Lin arrives at a set of methods comparing the formulaic language extracted automatically from a corpus of proficient formulaic language learners, and from native-speaker judgement of formulaic language in fluent speech from a native speaker of English. The majority of instances of formulaic language (but not all) in both cases aligned with intonation units, but this does not mean that intonation unit boundaries readily in and of themselves indicate formulaic language in the speech of formulaic language learners and native speakers; there are other factors to include if identification is to be successful. In the end, it appears that prosodic cues provide some but not all of the information needed to identify formulaic language in such situations.
Chapter 5 presents the second study, a profile of the intonation, stress and rhythm of formulaic language. Here, Lin examined the temporal and stress cues in the speech of a native speaker of British English, in an effort to test the notion that formulaic language has a distinctive prosody. While results were not particularly conclusive for measures such as articulation rate, and alignment with intonation unit boundaries showed little of note, the formulaic language in the speech of the participant did display much less stress than general speech. This piece of research takes us a step in the direction of identifying whether there is in fact a distinctive prosody of formulaic language.
Chapter 6 rounds out the three studies contained in the book. This study investigates whether playing original audio files for judges improves the strength of formulaic language judgement and inter-judge reliability. Thirty judges listened to audio recordings to identify formulaic language and also participated in a post-judgement interview, which revealed that they felt that the listening helped their judgements in various ways including providing a different perspective on the data, and heightened sensitivity to nuances of meaning in the recording. In the end, the data show that stronger judgements and greater inter-judge reliability did in fact result from listening to audio files of data.
Chapter 7 presents some conclusions about the prosody of formulaic language. Lin reviews the results and findings of the three studies and also reflects on the limitations of the methods employed as well as the fact that more issues have been raised by the results. The biggest issue is the long-standing one of identifying formulaic language, along with the problem of sampling formulaic language in data – if the samples are not valid, the research is heavily compromised. Lin calls for a multimodal approach to such research, involving the use of audio as well as written data, and calls for future research to employ this approach as well as using larger numbers of native-speaker judges, and reporting inter-rater reliability as a matter of course. She also presents some ideas for applications to teaching, including raising awareness by drawing learner attention to the prosodic cues in formulaic language.
This volume is a remarkable step forward in the study of formulaic language. It contains exhaustive literature reviews on a range of topic areas within the study of formulaic language, and is a go-to resource for up-to-date information on the prosody of formulaic language. More than this, however, the three studies contained in the book drive our knowledge forward, providing new information on native-speaker judgement, the prosodic profile of formulaic language in spontaneous speech, and the utility of a multimedia approach to data in judgement-focused research. Perhaps the greatest contribution here, however, is the model presented by the innovative methods employed in Lin's three studies, including use of a large group of judges, careful consideration of the components of prosody, and use of multiple data sources, among others. As the body of research on formulaic language grows in future, this volume will no doubt remain a standard or a classic in the field, certain to be a point of reference for all who conduct research into this elusive yet fundamental aspect of language.
